Save the kid, ruin the suit; Acceptable utility exchange rates; Distributed utility calculations; Civic duties matter
This is a follow up to my rambly incoherent comment on https://www.lesswrong.com/posts/cujpciCqNbawBihhQ/self-integrity-and-the-drowning-child
Let’s put some rough math on the drowning example
Statistical value of human life is $5,000,000.
Death of a child is a massive utility loss for parents and close relatives(several of them might even be willing to die in an attempt to save the child!)...
Death of a child is a utility hit of … $15,000,000 for various people. Rough number, but I don’t think grossly inflated, if anything probably too low.
A good suit is like $5000(?).
If you think there’s a 1% chance that taking 10 seconds to take off the suit, will kill the kid, you’re expected values are -$150,000 vs -$5,000.
Now. It’s -$150,000 of OTHER people’s utility vs $5000 of your own.
But if you take the suit saving option, this implies that your exchange rate of OTHER PEOPLE utility to PERSONAL utility is like 30 to 1. That’s pretty much an awful person, in my opinion. A literal car thief converts less than 2 OTHER for 1 PERSONAL.
Society should have some standards for what individual utility functions are acceptable
Claim 1.
We should expect people to care about other people, as in there should be SOME exchange rate between OTHER utility and PERSONAL utility.
Corollary
Individual preferences are off the table. Ah but what is individual?
We firewall people’s truly … individual preferences. You like oranges or pineapples, who cares. We only require that you care SOME AMOUNT about the people around you as well. Anything purely personal is no one else’s business.
But what is purely personal? Maybe the existence of a meat eater or a broccoli eater produces negative utility for me… whatever we come up with some rule based definition of personal, if your society ends up being too intrusive you’ll find out in the violent rebellion later on.
Claim 2.
The exchange rate should probably be (less than 2 OTHER) to 1 PERSONAL.
And even that’s high. Maybe like 1.1 OTHER to 1 PERSONAL.
Also there’s some tolerance to ignoring small utility inefficiencies. Like you should give money to the homeless on the pure local utility calculation, but it’s not that big a deal. Injure someone by speeding to not miss the first 10 minutes of a movie.… yeah throw them in the hole, officer.
Also 1 − 1 is probably … not computationally efficient. Like individuals know themselves best...
Actually. Maybe 1 − 1, but with an uncertainty factor, since I’m not 100% sure how much utility OTHER person is actually gaining or losing. Hmm, but then an overly considerate person might leave net utility on the table since they overestimate the utility cost to the other person. And blabla, this is a whole tangent of its own that’s probably somewhere in the sequences anyway so whatever.
Claim 3.
For computational efficiency purposes, an individual can legitimately cap the number of other people’s utilities directly integrated into their utility calculation, as long as they have a catch-all factor for ‘the general good’.
This is tricky and I’m still not happy with the explanation.
We can’t have everyone trying to maximise everyone else’s individual utility in a weird soup chaos computation, because you get stuck in recursive update loops and it’s just too complicated to even be aware of every other individual in the world and their detailed quirky preferences.
We need rules to properly distribute the computation, the responsibility and the costs to individuals. We’re all slightly responsible for X’s problem, but who is SO RESPONSIBLE THAT THEY MUST ACT for it? We’re all trying to maximise general utility but who does the buck stop with?
So some simple rules(not exhaustive, there’s tons more in practice to bound the computation).
By default everyone is responsible for themselves.
If we see someone in ‘unusual’ trouble, try to help them.
If we end up in trouble, we can expect other people try to help us, so don’t panic too much and don’t be too risk averse.
We certainly don’t expect people to sacrifice a HUGE amount of utility for someone else, but we call them heroes when they do(“So and so died after they went into a burning building to save their neighbour’s kids, what a wonderful person! Mayor is building them a statue in the town centre”).
Why are you not required to actively seek out people in trouble? What about the people in trouble over in the next village, or on the other side of the country? Or in a different country?
Well again, this would be all-consuming and society would grind to a halt with net collapse of utility for everyone. No time to farm, there’s a guy starving 300 miles away and I’m not yet starving myself.
The expectation is that problems are solved LOCALLY by other people in a given area.
Surprise government
Plus ACTUALY WE DO actively take part in helping ‘everyone’(in our state/nation/thing) by setting up a government, contributing to it and expecting that it smooths out utility disparities in a total utility improving way. (Success of said entities may vary)
Which is actually a nice segway into
Why should everyone do INSERT_CIVIC_DUTY?
After all the odds of your individual vote changing the election result is very low. Well, elections are giant distributed computational exercises. If Bob alone doesn’t vote it doesn’t really affect society, but if nobody votes it’s probably really bad out there.
It’s a tragedy of the commons effect. Civic duties exist to maintain the metaphorical structure of the society. If people are shirking from those duties, well, society is resilient and it keeps trucking for a while. If too many people shirk, it all starts breaking apart.
The correct thing to do is to put a price on not voting and redistribute the money to the people that do vote.
Amazingly, that’s literally what societies do! (Since voters determine government and government determines fiscal policy).
Now the complicated reality is that voting actually doesn’t matter that much, because sometimes none of the options on the ballot are going to fix INSERT_ISSUE. What you really need is complex engagement with the politics and policies of the day, so that you vote in an informed way and select actually good politicians. But this is a very complicated phenomenon to verify, whereas voting is easily legible, so at least vote.
And in fact, civic duty dictates that you actually get involved in politics if all the options on the table suck.
Obviously I’m shirking on that just like everyone I know, and that’s probably the self-reinforcing source of my political dissatisfaction. But hey, I also need to go on a diet and I just ordered a delicious pizza.
Key takeaway
“Voting doesn’t change anything” and pretty much all cynicism and inaction promoting memes are thought parasites designed to trick people into self-disenfranchising.
If your culture-tribe doesn’t vote, expect your life to get a lot worse over time due to government negligence or even maliciousness driven by your voting enemies. This can be forestalled by other societal effects(maybe your tribe is really employable so you do really well economically), but it remains a drag even if it’s overcome by those other effects.
Plus what happens when those other effects go away? Government is a more stable feature of our lives than a given instantiation of labour market structure.
Would you in reality choose to kill a sibling (or your child) to save 2 other unknown people? I personally would not kill my child even if it saved a million other people (the only circumstance where I would is to avoid human extinction or near-extinction). I’m pretty skeptical of anyone who would not choose this way, and I think they are not properly imagining what it would be like to kill a child. The actual exchange rate is much higher than 30 to 1 in this case.
Is the close relation actively killing people? I don’t think it’s an unreasonable standard to say you should attempt to kill your own child if they’re about to go on a spree killing and you definitely have exhausted all other options. I might fail it, but I’d definitely think I was a bad person for it(granted raising a spree killer in the first place is the bigger fault here).
How else are their lives trading so favourably? Organ transfers are pretty 1-1, tho maybe a more general policy encouraging people to donate spare kidneys wouldn’t be terrible. Also certainly people have encouraged their kids to go to war and so on, that’s a probabilistic sacrifice for the greater good.
Also killing/death tends to be a bit of a utility singularity so even the clumsy math of regular mode utilitarianism breaks down. How much utility do I lose by killing my child? Possibly infinite? Like I’m probably going to kill myself afterwards, surely that counts as a singularity.
Would I encourage myself or a relative to donate a kidney to save a life? Eh. Maybe, again there’s potential of death when donating an organ, so singularity type stuff slips in maybe.
Just because singularities exist in certain conditions of a theory doesn’t mean it’s unusable in finite number cases.
If you believe that you have an obligation to treat your own utility as worth around 1.1x that of other peoples’, and that applying a high ratio in this regard is monstrous, that seems to straightforwardly imply that you have e.g. an obligation to give away almost all of your income, and that many common and socially accepted human behaviours (such as failure to do so) are monstrous.
Do you endorse that implication?
I observe that there is geographic, genetic, and social discounting. So 1 personal utility is worth: .99 wife utility .99 child utility .9 parent who lives nearby utility .9 next door neighbor utility .75 friend 100 miles away .5 rando in same city .1 rando 1000 miles away in same country .01 rando in Africa
No because of the things I say in Claim 3. Like. If I were to do it alone, that would sort of be fine. But if everyone were to live that way, everyone would be miserable(something something Kant’s categorical imperative, what if everyone adopted this behaviour, would that work?).
I guess, there’s a difference between what is utility maximising for an individual to do in a given society, and what is a utility maximising way for individuals to behave in an ideal society.
Like society should be such that Claim 3 is all you need, localized responsibility + government redistribution.
If everyone donated income to everyone who needed it at least 10% more (in utility/$ terms), then the equilibrium would be a state where almost everyone gets to keep almost all their income because everyone in bad but fixable situations is now doing a lot better and there are no major utility gradients left. There would still be people who are worse off, but they’re worse off in ways that can’t be easily remedied by things that money can buy. So no, Kant doesn’t have any objections.
So if this is the requirement to be ethical, then nearly everyone in the world is unethical. Which isn’t a surprising conclusion, but it’s surprising for someone to both say “this is ethically necessary” and “no I won’t do that”.
this would be much closer to the Pareto frontier then our curren social organization! unfortunately, this is NOT how society work. if you will operate like that, you will loss almost all your resources.
but it’s more complicated then that—why not gate this on cooperation? why should i give 1 dollar for 2 of someone else dollars, when they will not do that for me?
and this is why all this scheme doesn’t work. every such plan need to account for defectors, and it doesn’t look like you address it, anywhere.
on the issue of politics—most people who involve in politics make things worse. before declaring that it’s people duty to do something, it’s important to verify this is net-positive thing to do. if i look on people involved in politics and decide that less politics would have been better to society, then my duty is to NOT get involved in politics. or at least, not to get involved more then the level that i believe is the right level of involvement.
but… i really don’t see how all this politics even connected to the first half of the post, about the right ratio of my utility : other person utility?
Good points! I didn’t get into ‘How do you really calculate the net-societal utility outcome of your actions, including second+ order effects?’ since I think even the 1st order immediate consequence calculation is intractable.
In practice you shouldn’t … help people that will do harm with your help. I think this is one of the limits of pacifism(with Kaladin’s dad being an example of that, incidentally) at some point passive obedience to an unjust opressor has the same consequences for the other people they harm as active cooperation. It is a moral duty to actually do something to harm or at least minimally help a person or institution or government from doing bad things. Just saying I don’t like what this Hitler guy is doing with my tax money isn’t really acceptable once the harm he’s causing becomes really monstrous. So if I’m really calculating paying 2 personal utility to generate 4 utility for Bob, I should take into account what Bob’s actually gonna do with his 4 utility. But again that becomes computationally impossible almost immediately, hence we use heuristics(aka moral principles) to dictate how we should behave.
By politics I mean governance, collective action via voluntary organisations and state action.
The connection seems clear to me: I want to pool resources with others in a way that makes all of us better off. Increasingly elaborate and large scale systems of social coordination is how we do that and that’s what modern states are. (for better or for worse and as hijacked by niche, elite special interests as they can seem to be and/or actually are)
As … disappointing as contemporary western governments are, I still think most ‘charity’ or utility redistribution in modern societies is done by government via schools, healthcare, pensions, police and other security systems and relatively cheap/free infrastructure. These are all things that were privileges or luxuries in the past that are now baseline and we all pay for them together.
The modern idea is that politics is dirty and gross. And pretty much any politician I can think of off the top of my head is at best disappointing, at worst vile. However, developed societies went from feudal serfdom or slavery or highly unequal large underclass early industrial society to modern social democratic welfare states with a historically relatively high standard of living even for the worst off(or at least for the almost worst off, the really lumpenproletariat among us aren’t doing that great, but the people at the bottom 15% threshold are, relatively speaking).
This transformation happened because people, not necessarily professional politicians but some were that as well, pushed for change in an intentional, organised and persistent way. And they got it.
The grossness of modern politicians is a problem that will either be solved by better politicians emerging or will destroy our societies. Crap elites kill civilizations.
Without organised, collective action towards the goal of improving our lives in specific ways, with specific policies … we won’t get the things we want. Society doesn’t get better randomly, it gets better because groups of people agitate in a direction they think will make it better and sometimes they get what they want and sometimes what they wanted actually was a good idea.